On Thu, Dec 4, 2014 at 10:17 AM, Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl> wrote:
On Thu, Dec 04, 2014 at 09:39:35AM -0500, Matthew Miller wrote:
> [tick tock] would mean alternating between concentrating on release
> features and on release engineering and QA process and tooling. During
> the "tick", we'd focus on new features and minimize unrelated rel-eng
> change. During the "tock", we'd focus on the tools, and minimize change
> that might affect that.
This view is very engineering- and internal-project- centric. If you
only mean tools developed in Fedora and for Fedora, please say so.

If you mean everything, then this plan doesn't sound good. Delaying
features for a couple of months works with anaconda, but not with any
independent external project which is going to accumulate features
irrespectively of what Fedora is doing.
In case of large projects this is bad for two reasons:
1. a jump twice as big when it is finally taken
2. in the "tock" release, wasted effort is spent on backporting fixes
   while upstream moves on.

And of course Fedora is not "First" anymore.

I can see how this kind of cadence might be useful for some projects
or packages, but they should decide on their own, independently.


As a user, I love the idea of stabilizing Fedora tools on every other release. As a developer and a user, I like the idea of those releases labeled differently, so that individual packages can adopt those semantics if it makes sense for them, independently. Basically, the name or release "type" (not "LTS", but whatever label is appropriate), could be a semantic hook for packages to follow or not.

--
Christopher L Tubbs II
http://gravatar.com/ctubbsii